Scaffolding Learning With Pedagogical Agents in Advanced Learning Technologies: Understanding the Role of Self- Versus External Regulation

• In Event: 75.019 - Scaffolding Students' Learning in Formal and Informal Contexts

Abstract

Objectives: Self-regulated learning (SRL) involves actively monitoring and regulating key cognitive, affective, metacognitive, and motivational (CAMM) processes during complex science learning with advanced learning technologies (e.g., multi-agent systems) (Azevedo & Aleven, 2013). However, most learners do not have the skills to effectively and accurately enact these processes during learning and therefore fail to demonstrate increased understanding of complex science topics (Winne & Azevedo, in press). Researchers have addressed this issue by using pedagogical agents (PAs) to scaffold students’ challenges in monitoring and regulating their own learning about complex science topics (Azevedo et al., 2012; Lester et al., 2013). Specifically, we have converged product data (pretest and posttest scores) with several on-line measures of CAMM processes (e.g., log-files, metacognitive judgments, use of SRL palette, and facial expressions of emotions) to examine the effectiveness of self- and externally-regulated learning with (adaptive) or without the agents (non-adaptive) during learning with MetaTutor (Azevedo et al., 2013). In this study, we focus on presenting empirical evidence, using process and product data, regarding the effectiveness of MetaTutor’s four PAs’ scaffolding of students learning about a complex science topic.

Methods and Results: One hundred fifty (N = 150) university students took part in a two-day experiment with MetaTutor (Azevedo et al., 2010, 2013) to learn about the human circulatory system. They were instructed to use several CAMM SRL processes during their learning session (e.g., setting-relevant learning goals, metacognitive monitoring, effective learning strategies). Participants were randomly assigned either to the adaptive or non-adaptive condition. The effectiveness of PAs’ SRL scaffolding was assessed based on analyses of participants’ two-hour session with MetaTutor where we collected from each participant: eye-tracking, video recording of the face (for affect detection and classification), log-files (e.g., quiz results, summaries and metacognitive judgments, learner-agent dialogue), and physiological data. We collected pretest and posttest data and several self-report measures on agent likeability. Our results indicated statistically significant differences between MetaTutor conditions. Learners in the adaptive scaffolding condition had higher learning gains, sub-goal quiz scores, spent more time engaging in each learning sub-goal, engaged in more help-seeking behavior, inspected more relevant multimedia materials during learning, deployed more sophisticated learning strategies, and made higher metacognitive judgments, compared to participants in the non-adaptive condition. These results suggest the importance of SRL scaffolding and the promotion of effective SRL strategies while interacting with intelligent, agent-based hypermedia-learning environments in learning about complex topics.

Significance: Research on scaffolding has focused on assessing the effectiveness of self- versus externally regulated learning (Järvelä & Hadwin, 2013) Understanding the effectiveness of individual learners’ versus externally-regulating agent’s (i.e., PAs’) dynamic enactment of SRL processes during scaffolding is the key to enhancing complex (STEM) learning and performance. The multi-channel data sources provide evidence that has the potential to advance current conceptual, theoretical, methodological, and analytical frameworks related to scaffolding and SRL, based on self- and external-regulation (Molenaar & Järvelä, in press). These advances will allow researchers to design more effective PAs that are capable of providing adaptive scaffolding that is responsive to students’ CAMM processes during learning.

Authors

• Michelle Taub, North Carolina State University

• Nicholas Vincent Mudrick, North Carolina State University

• Roger Azevedo, North Carolina State University